Upload folder using huggingface_hub
Browse files- config.json +290 -0
- generation_config.json +270 -0
- model-00001-of-00002.safetensors +3 -0
- model-00002-of-00002.safetensors +3 -0
- model.safetensors.index.json +0 -0
- optimizer.pt +3 -0
- preprocessor_config.json +14 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +758 -0
- training_args.bin +3 -0
config.json
ADDED
@@ -0,0 +1,290 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "NbAiLab/nb-whisper-large-verbatim",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"activation_function": "gelu",
|
5 |
+
"alignment_heads": [
|
6 |
+
[
|
7 |
+
7,
|
8 |
+
0
|
9 |
+
],
|
10 |
+
[
|
11 |
+
10,
|
12 |
+
17
|
13 |
+
],
|
14 |
+
[
|
15 |
+
12,
|
16 |
+
18
|
17 |
+
],
|
18 |
+
[
|
19 |
+
13,
|
20 |
+
12
|
21 |
+
],
|
22 |
+
[
|
23 |
+
16,
|
24 |
+
1
|
25 |
+
],
|
26 |
+
[
|
27 |
+
17,
|
28 |
+
14
|
29 |
+
],
|
30 |
+
[
|
31 |
+
19,
|
32 |
+
11
|
33 |
+
],
|
34 |
+
[
|
35 |
+
21,
|
36 |
+
4
|
37 |
+
],
|
38 |
+
[
|
39 |
+
24,
|
40 |
+
1
|
41 |
+
],
|
42 |
+
[
|
43 |
+
25,
|
44 |
+
6
|
45 |
+
]
|
46 |
+
],
|
47 |
+
"apply_spec_augment": false,
|
48 |
+
"architectures": [
|
49 |
+
"WhisperForConditionalGeneration"
|
50 |
+
],
|
51 |
+
"attention_dropout": 0,
|
52 |
+
"begin_suppress_tokens": [
|
53 |
+
220,
|
54 |
+
50257
|
55 |
+
],
|
56 |
+
"bos_token_id": 50257,
|
57 |
+
"classifier_proj_size": 256,
|
58 |
+
"d_model": 1280,
|
59 |
+
"decoder_attention_heads": 20,
|
60 |
+
"decoder_ffn_dim": 5120,
|
61 |
+
"decoder_layerdrop": 0,
|
62 |
+
"decoder_layers": 32,
|
63 |
+
"decoder_start_token_id": 50258,
|
64 |
+
"dropout": 0,
|
65 |
+
"encoder_attention_heads": 20,
|
66 |
+
"encoder_ffn_dim": 5120,
|
67 |
+
"encoder_layerdrop": 0,
|
68 |
+
"encoder_layers": 32,
|
69 |
+
"eos_token_id": 50257,
|
70 |
+
"forced_decoder_ids": null,
|
71 |
+
"init_std": 0.02,
|
72 |
+
"is_encoder_decoder": true,
|
73 |
+
"lang_ids": [
|
74 |
+
50259,
|
75 |
+
50260,
|
76 |
+
50261,
|
77 |
+
50262,
|
78 |
+
50263,
|
79 |
+
50264,
|
80 |
+
50265,
|
81 |
+
50266,
|
82 |
+
50267,
|
83 |
+
50268,
|
84 |
+
50269,
|
85 |
+
50270,
|
86 |
+
50271,
|
87 |
+
50272,
|
88 |
+
50273,
|
89 |
+
50274,
|
90 |
+
50275,
|
91 |
+
50276,
|
92 |
+
50277,
|
93 |
+
50278,
|
94 |
+
50279,
|
95 |
+
50280,
|
96 |
+
50281,
|
97 |
+
50282,
|
98 |
+
50283,
|
99 |
+
50284,
|
100 |
+
50285,
|
101 |
+
50286,
|
102 |
+
50287,
|
103 |
+
50288,
|
104 |
+
50289,
|
105 |
+
50290,
|
106 |
+
50291,
|
107 |
+
50292,
|
108 |
+
50293,
|
109 |
+
50294,
|
110 |
+
50295,
|
111 |
+
50296,
|
112 |
+
50297,
|
113 |
+
50298,
|
114 |
+
50299,
|
115 |
+
50300,
|
116 |
+
50301,
|
117 |
+
50302,
|
118 |
+
50303,
|
119 |
+
50304,
|
120 |
+
50305,
|
121 |
+
50306,
|
122 |
+
50307,
|
123 |
+
50308,
|
124 |
+
50309,
|
125 |
+
50310,
|
126 |
+
50311,
|
127 |
+
50312,
|
128 |
+
50313,
|
129 |
+
50314,
|
130 |
+
50315,
|
131 |
+
50316,
|
132 |
+
50317,
|
133 |
+
50318,
|
134 |
+
50319,
|
135 |
+
50320,
|
136 |
+
50321,
|
137 |
+
50322,
|
138 |
+
50323,
|
139 |
+
50324,
|
140 |
+
50325,
|
141 |
+
50326,
|
142 |
+
50327,
|
143 |
+
50328,
|
144 |
+
50329,
|
145 |
+
50330,
|
146 |
+
50331,
|
147 |
+
50332,
|
148 |
+
50333,
|
149 |
+
50334,
|
150 |
+
50335,
|
151 |
+
50336,
|
152 |
+
50337,
|
153 |
+
50338,
|
154 |
+
50339,
|
155 |
+
50340,
|
156 |
+
50341,
|
157 |
+
50342,
|
158 |
+
50343,
|
159 |
+
50344,
|
160 |
+
50345,
|
161 |
+
50346,
|
162 |
+
50347,
|
163 |
+
50348,
|
164 |
+
50349,
|
165 |
+
50350,
|
166 |
+
50351,
|
167 |
+
50352,
|
168 |
+
50353,
|
169 |
+
50354,
|
170 |
+
50355,
|
171 |
+
50356,
|
172 |
+
50357,
|
173 |
+
50358
|
174 |
+
],
|
175 |
+
"mask_feature_length": 10,
|
176 |
+
"mask_feature_min_masks": 0,
|
177 |
+
"mask_feature_prob": 0,
|
178 |
+
"mask_time_length": 10,
|
179 |
+
"mask_time_min_masks": 2,
|
180 |
+
"mask_time_prob": 0.05,
|
181 |
+
"max_length": 448,
|
182 |
+
"max_source_positions": 1500,
|
183 |
+
"max_target_positions": 448,
|
184 |
+
"median_filter_width": 7,
|
185 |
+
"model_type": "whisper",
|
186 |
+
"num_hidden_layers": 32,
|
187 |
+
"num_mel_bins": 128,
|
188 |
+
"pad_token_id": 50256,
|
189 |
+
"scale_embedding": false,
|
190 |
+
"suppress_ids": [
|
191 |
+
1,
|
192 |
+
2,
|
193 |
+
7,
|
194 |
+
8,
|
195 |
+
9,
|
196 |
+
10,
|
197 |
+
14,
|
198 |
+
25,
|
199 |
+
26,
|
200 |
+
27,
|
201 |
+
28,
|
202 |
+
29,
|
203 |
+
31,
|
204 |
+
58,
|
205 |
+
59,
|
206 |
+
60,
|
207 |
+
61,
|
208 |
+
62,
|
209 |
+
63,
|
210 |
+
90,
|
211 |
+
91,
|
212 |
+
92,
|
213 |
+
93,
|
214 |
+
359,
|
215 |
+
503,
|
216 |
+
522,
|
217 |
+
542,
|
218 |
+
873,
|
219 |
+
893,
|
220 |
+
902,
|
221 |
+
918,
|
222 |
+
922,
|
223 |
+
931,
|
224 |
+
1350,
|
225 |
+
1853,
|
226 |
+
1982,
|
227 |
+
2460,
|
228 |
+
2627,
|
229 |
+
3246,
|
230 |
+
3253,
|
231 |
+
3268,
|
232 |
+
3536,
|
233 |
+
3846,
|
234 |
+
3961,
|
235 |
+
4183,
|
236 |
+
4667,
|
237 |
+
6585,
|
238 |
+
6647,
|
239 |
+
7273,
|
240 |
+
9061,
|
241 |
+
9383,
|
242 |
+
10428,
|
243 |
+
10929,
|
244 |
+
11938,
|
245 |
+
12033,
|
246 |
+
12331,
|
247 |
+
12562,
|
248 |
+
13793,
|
249 |
+
14157,
|
250 |
+
14635,
|
251 |
+
15265,
|
252 |
+
15618,
|
253 |
+
16553,
|
254 |
+
16604,
|
255 |
+
18362,
|
256 |
+
18956,
|
257 |
+
20075,
|
258 |
+
21675,
|
259 |
+
22520,
|
260 |
+
26130,
|
261 |
+
26161,
|
262 |
+
26435,
|
263 |
+
28279,
|
264 |
+
29464,
|
265 |
+
31650,
|
266 |
+
32302,
|
267 |
+
32470,
|
268 |
+
36865,
|
269 |
+
42863,
|
270 |
+
47425,
|
271 |
+
49870,
|
272 |
+
50254,
|
273 |
+
50258,
|
274 |
+
50359,
|
275 |
+
50360,
|
276 |
+
50361,
|
277 |
+
50362,
|
278 |
+
50363
|
279 |
+
],
|
280 |
+
"suppress_ids_begin": [
|
281 |
+
220,
|
282 |
+
50257
|
283 |
+
],
|
284 |
+
"suppress_tokens": [],
|
285 |
+
"torch_dtype": "float32",
|
286 |
+
"transformers_version": "4.38.1",
|
287 |
+
"use_cache": true,
|
288 |
+
"use_weighted_layer_sum": false,
|
289 |
+
"vocab_size": 51866
|
290 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,270 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alignment_heads": [
|
3 |
+
[
|
4 |
+
7,
|
5 |
+
0
|
6 |
+
],
|
7 |
+
[
|
8 |
+
10,
|
9 |
+
17
|
10 |
+
],
|
11 |
+
[
|
12 |
+
12,
|
13 |
+
18
|
14 |
+
],
|
15 |
+
[
|
16 |
+
13,
|
17 |
+
12
|
18 |
+
],
|
19 |
+
[
|
20 |
+
16,
|
21 |
+
1
|
22 |
+
],
|
23 |
+
[
|
24 |
+
17,
|
25 |
+
14
|
26 |
+
],
|
27 |
+
[
|
28 |
+
19,
|
29 |
+
11
|
30 |
+
],
|
31 |
+
[
|
32 |
+
21,
|
33 |
+
4
|
34 |
+
],
|
35 |
+
[
|
36 |
+
24,
|
37 |
+
1
|
38 |
+
],
|
39 |
+
[
|
40 |
+
25,
|
41 |
+
6
|
42 |
+
]
|
43 |
+
],
|
44 |
+
"begin_suppress_tokens": [
|
45 |
+
220,
|
46 |
+
50257
|
47 |
+
],
|
48 |
+
"bos_token_id": 50257,
|
49 |
+
"decoder_start_token_id": 50258,
|
50 |
+
"eos_token_id": 50257,
|
51 |
+
"forced_decoder_ids": [
|
52 |
+
[
|
53 |
+
1,
|
54 |
+
50288
|
55 |
+
],
|
56 |
+
[
|
57 |
+
2,
|
58 |
+
50360
|
59 |
+
],
|
60 |
+
[
|
61 |
+
3,
|
62 |
+
50364
|
63 |
+
]
|
64 |
+
],
|
65 |
+
"is_multilingual": true,
|
66 |
+
"lang_to_id": {
|
67 |
+
"<|af|>": 50327,
|
68 |
+
"<|am|>": 50334,
|
69 |
+
"<|ar|>": 50272,
|
70 |
+
"<|as|>": 50350,
|
71 |
+
"<|az|>": 50304,
|
72 |
+
"<|ba|>": 50355,
|
73 |
+
"<|be|>": 50330,
|
74 |
+
"<|bg|>": 50292,
|
75 |
+
"<|bn|>": 50302,
|
76 |
+
"<|bo|>": 50347,
|
77 |
+
"<|br|>": 50309,
|
78 |
+
"<|bs|>": 50315,
|
79 |
+
"<|ca|>": 50270,
|
80 |
+
"<|cs|>": 50283,
|
81 |
+
"<|cy|>": 50297,
|
82 |
+
"<|da|>": 50285,
|
83 |
+
"<|de|>": 50261,
|
84 |
+
"<|el|>": 50281,
|
85 |
+
"<|en|>": 50259,
|
86 |
+
"<|es|>": 50262,
|
87 |
+
"<|et|>": 50307,
|
88 |
+
"<|eu|>": 50310,
|
89 |
+
"<|fa|>": 50300,
|
90 |
+
"<|fi|>": 50277,
|
91 |
+
"<|fo|>": 50338,
|
92 |
+
"<|fr|>": 50265,
|
93 |
+
"<|gl|>": 50319,
|
94 |
+
"<|gu|>": 50333,
|
95 |
+
"<|haw|>": 50352,
|
96 |
+
"<|ha|>": 50354,
|
97 |
+
"<|he|>": 50279,
|
98 |
+
"<|hi|>": 50276,
|
99 |
+
"<|hr|>": 50291,
|
100 |
+
"<|ht|>": 50339,
|
101 |
+
"<|hu|>": 50286,
|
102 |
+
"<|hy|>": 50312,
|
103 |
+
"<|id|>": 50275,
|
104 |
+
"<|is|>": 50311,
|
105 |
+
"<|it|>": 50274,
|
106 |
+
"<|ja|>": 50266,
|
107 |
+
"<|jw|>": 50356,
|
108 |
+
"<|ka|>": 50329,
|
109 |
+
"<|kk|>": 50316,
|
110 |
+
"<|km|>": 50323,
|
111 |
+
"<|kn|>": 50306,
|
112 |
+
"<|ko|>": 50264,
|
113 |
+
"<|la|>": 50294,
|
114 |
+
"<|lb|>": 50345,
|
115 |
+
"<|ln|>": 50353,
|
116 |
+
"<|lo|>": 50336,
|
117 |
+
"<|lt|>": 50293,
|
118 |
+
"<|lv|>": 50301,
|
119 |
+
"<|mg|>": 50349,
|
120 |
+
"<|mi|>": 50295,
|
121 |
+
"<|mk|>": 50308,
|
122 |
+
"<|ml|>": 50296,
|
123 |
+
"<|mn|>": 50314,
|
124 |
+
"<|mr|>": 50320,
|
125 |
+
"<|ms|>": 50282,
|
126 |
+
"<|mt|>": 50343,
|
127 |
+
"<|my|>": 50346,
|
128 |
+
"<|ne|>": 50313,
|
129 |
+
"<|nl|>": 50271,
|
130 |
+
"<|nn|>": 50342,
|
131 |
+
"<|no|>": 50288,
|
132 |
+
"<|oc|>": 50328,
|
133 |
+
"<|pa|>": 50321,
|
134 |
+
"<|pl|>": 50269,
|
135 |
+
"<|ps|>": 50340,
|
136 |
+
"<|pt|>": 50267,
|
137 |
+
"<|ro|>": 50284,
|
138 |
+
"<|ru|>": 50263,
|
139 |
+
"<|sa|>": 50344,
|
140 |
+
"<|sd|>": 50332,
|
141 |
+
"<|si|>": 50322,
|
142 |
+
"<|sk|>": 50298,
|
143 |
+
"<|sl|>": 50305,
|
144 |
+
"<|sn|>": 50324,
|
145 |
+
"<|so|>": 50326,
|
146 |
+
"<|sq|>": 50317,
|
147 |
+
"<|sr|>": 50303,
|
148 |
+
"<|su|>": 50357,
|
149 |
+
"<|sv|>": 50273,
|
150 |
+
"<|sw|>": 50318,
|
151 |
+
"<|ta|>": 50287,
|
152 |
+
"<|te|>": 50299,
|
153 |
+
"<|tg|>": 50331,
|
154 |
+
"<|th|>": 50289,
|
155 |
+
"<|tk|>": 50341,
|
156 |
+
"<|tl|>": 50348,
|
157 |
+
"<|tr|>": 50268,
|
158 |
+
"<|tt|>": 50351,
|
159 |
+
"<|uk|>": 50280,
|
160 |
+
"<|ur|>": 50290,
|
161 |
+
"<|uz|>": 50337,
|
162 |
+
"<|vi|>": 50278,
|
163 |
+
"<|yi|>": 50335,
|
164 |
+
"<|yo|>": 50325,
|
165 |
+
"<|yue|>": 50358,
|
166 |
+
"<|zh|>": 50260
|
167 |
+
},
|
168 |
+
"language": "<|no|>",
|
169 |
+
"max_initial_timestamp_index": 1,
|
170 |
+
"max_length": 448,
|
171 |
+
"no_timestamps_token_id": 50364,
|
172 |
+
"pad_token_id": 50257,
|
173 |
+
"return_timestamps": false,
|
174 |
+
"suppress_tokens": [
|
175 |
+
1,
|
176 |
+
2,
|
177 |
+
7,
|
178 |
+
8,
|
179 |
+
9,
|
180 |
+
10,
|
181 |
+
14,
|
182 |
+
25,
|
183 |
+
26,
|
184 |
+
27,
|
185 |
+
28,
|
186 |
+
29,
|
187 |
+
31,
|
188 |
+
58,
|
189 |
+
59,
|
190 |
+
60,
|
191 |
+
61,
|
192 |
+
62,
|
193 |
+
63,
|
194 |
+
90,
|
195 |
+
91,
|
196 |
+
92,
|
197 |
+
93,
|
198 |
+
359,
|
199 |
+
503,
|
200 |
+
522,
|
201 |
+
542,
|
202 |
+
873,
|
203 |
+
893,
|
204 |
+
902,
|
205 |
+
918,
|
206 |
+
922,
|
207 |
+
931,
|
208 |
+
1350,
|
209 |
+
1853,
|
210 |
+
1982,
|
211 |
+
2460,
|
212 |
+
2627,
|
213 |
+
3246,
|
214 |
+
3253,
|
215 |
+
3268,
|
216 |
+
3536,
|
217 |
+
3846,
|
218 |
+
3961,
|
219 |
+
4183,
|
220 |
+
4667,
|
221 |
+
6585,
|
222 |
+
6647,
|
223 |
+
7273,
|
224 |
+
9061,
|
225 |
+
9383,
|
226 |
+
10428,
|
227 |
+
10929,
|
228 |
+
11938,
|
229 |
+
12033,
|
230 |
+
12331,
|
231 |
+
12562,
|
232 |
+
13793,
|
233 |
+
14157,
|
234 |
+
14635,
|
235 |
+
15265,
|
236 |
+
15618,
|
237 |
+
16553,
|
238 |
+
16604,
|
239 |
+
18362,
|
240 |
+
18956,
|
241 |
+
20075,
|
242 |
+
21675,
|
243 |
+
22520,
|
244 |
+
26130,
|
245 |
+
26161,
|
246 |
+
26435,
|
247 |
+
28279,
|
248 |
+
29464,
|
249 |
+
31650,
|
250 |
+
32302,
|
251 |
+
32470,
|
252 |
+
36865,
|
253 |
+
42863,
|
254 |
+
47425,
|
255 |
+
49870,
|
256 |
+
50254,
|
257 |
+
50258,
|
258 |
+
50359,
|
259 |
+
50360,
|
260 |
+
50361,
|
261 |
+
50362,
|
262 |
+
50363
|
263 |
+
],
|
264 |
+
"task": "transcribe",
|
265 |
+
"task_to_id": {
|
266 |
+
"transcribe": 50360,
|
267 |
+
"translate": 50359
|
268 |
+
},
|
269 |
+
"transformers_version": "4.38.1"
|
270 |
+
}
|
model-00001-of-00002.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4c838a74f852dcf1a5f5c4f06252143c9279d695963f6c6ef54a4c770d65b748
|
3 |
+
size 4993448880
|
model-00002-of-00002.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:59f064da08c50ac29e5290657564024a4a59ffc39d6a0ae82ad33e9f3f958109
|
3 |
+
size 1180663192
|
model.safetensors.index.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0ccc7ab042538232bc6a996c863ec1ad1318d59f03b8c681586b24f26aef85a4
|
3 |
+
size 12333660476
|
preprocessor_config.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"chunk_length": 30,
|
3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
4 |
+
"feature_size": 128,
|
5 |
+
"hop_length": 160,
|
6 |
+
"n_fft": 400,
|
7 |
+
"n_samples": 480000,
|
8 |
+
"nb_max_frames": 3000,
|
9 |
+
"padding_side": "right",
|
10 |
+
"padding_value": 0.0,
|
11 |
+
"processor_class": "WhisperProcessor",
|
12 |
+
"return_attention_mask": false,
|
13 |
+
"sampling_rate": 16000
|
14 |
+
}
|
rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8bd57c654092f134895f518f4c8a68e3b53b679bfa87212cac8eb67a6ac2794a
|
3 |
+
size 14244
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:da423a39ed9647de8be7df65b0cb7b2806b048deca302384c81aa13270200849
|
3 |
+
size 1064
|
trainer_state.json
ADDED
@@ -0,0 +1,758 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 8.647140864714087,
|
3 |
+
"best_model_checkpoint": "/cluster/home/torstefl/Master/saved_model/Whisper/NB/checkpoint-546",
|
4 |
+
"epoch": 6.204545454545454,
|
5 |
+
"eval_steps": 21,
|
6 |
+
"global_step": 546,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.24,
|
13 |
+
"eval_rundkast_loss": 1.4755723476409912,
|
14 |
+
"eval_rundkast_runtime": 265.7554,
|
15 |
+
"eval_rundkast_samples_per_second": 0.963,
|
16 |
+
"eval_rundkast_steps_per_second": 0.015,
|
17 |
+
"eval_rundkast_wer": 17.823479005998287,
|
18 |
+
"step": 21
|
19 |
+
},
|
20 |
+
{
|
21 |
+
"epoch": 0.24,
|
22 |
+
"eval_nb_samtale_loss": 1.2217341661453247,
|
23 |
+
"eval_nb_samtale_runtime": 132.2755,
|
24 |
+
"eval_nb_samtale_samples_per_second": 0.968,
|
25 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
26 |
+
"eval_nb_samtale_wer": 14.504881450488144,
|
27 |
+
"step": 21
|
28 |
+
},
|
29 |
+
{
|
30 |
+
"epoch": 0.24,
|
31 |
+
"eval_bigbrother_loss": 1.8927885293960571,
|
32 |
+
"eval_bigbrother_runtime": 256.5203,
|
33 |
+
"eval_bigbrother_samples_per_second": 0.998,
|
34 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
35 |
+
"eval_bigbrother_wer": 29.599999999999998,
|
36 |
+
"step": 21
|
37 |
+
},
|
38 |
+
{
|
39 |
+
"epoch": 0.48,
|
40 |
+
"eval_rundkast_loss": 1.2454240322113037,
|
41 |
+
"eval_rundkast_runtime": 255.1947,
|
42 |
+
"eval_rundkast_samples_per_second": 1.003,
|
43 |
+
"eval_rundkast_steps_per_second": 0.016,
|
44 |
+
"eval_rundkast_wer": 17.13796058269066,
|
45 |
+
"step": 42
|
46 |
+
},
|
47 |
+
{
|
48 |
+
"epoch": 0.48,
|
49 |
+
"eval_nb_samtale_loss": 0.9530782103538513,
|
50 |
+
"eval_nb_samtale_runtime": 132.01,
|
51 |
+
"eval_nb_samtale_samples_per_second": 0.97,
|
52 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
53 |
+
"eval_nb_samtale_wer": 14.714086471408647,
|
54 |
+
"step": 42
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 0.48,
|
58 |
+
"eval_bigbrother_loss": 1.7297923564910889,
|
59 |
+
"eval_bigbrother_runtime": 255.4086,
|
60 |
+
"eval_bigbrother_samples_per_second": 1.002,
|
61 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
62 |
+
"eval_bigbrother_wer": 29.15555555555556,
|
63 |
+
"step": 42
|
64 |
+
},
|
65 |
+
{
|
66 |
+
"epoch": 0.72,
|
67 |
+
"eval_rundkast_loss": 0.9745193719863892,
|
68 |
+
"eval_rundkast_runtime": 255.3218,
|
69 |
+
"eval_rundkast_samples_per_second": 1.003,
|
70 |
+
"eval_rundkast_steps_per_second": 0.016,
|
71 |
+
"eval_rundkast_wer": 16.36675235646958,
|
72 |
+
"step": 63
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.72,
|
76 |
+
"eval_nb_samtale_loss": 0.7309895157814026,
|
77 |
+
"eval_nb_samtale_runtime": 131.5521,
|
78 |
+
"eval_nb_samtale_samples_per_second": 0.973,
|
79 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
80 |
+
"eval_nb_samtale_wer": 13.772663877266389,
|
81 |
+
"step": 63
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 0.72,
|
85 |
+
"eval_bigbrother_loss": 1.4885540008544922,
|
86 |
+
"eval_bigbrother_runtime": 255.8175,
|
87 |
+
"eval_bigbrother_samples_per_second": 1.001,
|
88 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
89 |
+
"eval_bigbrother_wer": 29.15555555555556,
|
90 |
+
"step": 63
|
91 |
+
},
|
92 |
+
{
|
93 |
+
"epoch": 0.95,
|
94 |
+
"eval_rundkast_loss": 0.6675350666046143,
|
95 |
+
"eval_rundkast_runtime": 258.7982,
|
96 |
+
"eval_rundkast_samples_per_second": 0.989,
|
97 |
+
"eval_rundkast_steps_per_second": 0.015,
|
98 |
+
"eval_rundkast_wer": 15.295629820051415,
|
99 |
+
"step": 84
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.95,
|
103 |
+
"eval_nb_samtale_loss": 0.5071679353713989,
|
104 |
+
"eval_nb_samtale_runtime": 132.2238,
|
105 |
+
"eval_nb_samtale_samples_per_second": 0.968,
|
106 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
107 |
+
"eval_nb_samtale_wer": 13.319386331938631,
|
108 |
+
"step": 84
|
109 |
+
},
|
110 |
+
{
|
111 |
+
"epoch": 0.95,
|
112 |
+
"eval_bigbrother_loss": 1.1185076236724854,
|
113 |
+
"eval_bigbrother_runtime": 255.4352,
|
114 |
+
"eval_bigbrother_samples_per_second": 1.002,
|
115 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
116 |
+
"eval_bigbrother_wer": 29.333333333333332,
|
117 |
+
"step": 84
|
118 |
+
},
|
119 |
+
{
|
120 |
+
"epoch": 1.14,
|
121 |
+
"grad_norm": 2.3595402240753174,
|
122 |
+
"learning_rate": 1.0000000000000002e-06,
|
123 |
+
"loss": 0.8426,
|
124 |
+
"step": 100
|
125 |
+
},
|
126 |
+
{
|
127 |
+
"epoch": 1.19,
|
128 |
+
"eval_rundkast_loss": 0.4894542694091797,
|
129 |
+
"eval_rundkast_runtime": 257.5365,
|
130 |
+
"eval_rundkast_samples_per_second": 0.994,
|
131 |
+
"eval_rundkast_steps_per_second": 0.016,
|
132 |
+
"eval_rundkast_wer": 18.466152527849186,
|
133 |
+
"step": 105
|
134 |
+
},
|
135 |
+
{
|
136 |
+
"epoch": 1.19,
|
137 |
+
"eval_nb_samtale_loss": 0.39404886960983276,
|
138 |
+
"eval_nb_samtale_runtime": 132.7392,
|
139 |
+
"eval_nb_samtale_samples_per_second": 0.964,
|
140 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
141 |
+
"eval_nb_samtale_wer": 12.377963737796374,
|
142 |
+
"step": 105
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 1.19,
|
146 |
+
"eval_bigbrother_loss": 0.8832881450653076,
|
147 |
+
"eval_bigbrother_runtime": 255.7717,
|
148 |
+
"eval_bigbrother_samples_per_second": 1.001,
|
149 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
150 |
+
"eval_bigbrother_wer": 29.42222222222222,
|
151 |
+
"step": 105
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 1.43,
|
155 |
+
"eval_rundkast_loss": 0.41296815872192383,
|
156 |
+
"eval_rundkast_runtime": 266.8919,
|
157 |
+
"eval_rundkast_samples_per_second": 0.959,
|
158 |
+
"eval_rundkast_steps_per_second": 0.015,
|
159 |
+
"eval_rundkast_wer": 17.13796058269066,
|
160 |
+
"step": 126
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 1.43,
|
164 |
+
"eval_nb_samtale_loss": 0.3332406282424927,
|
165 |
+
"eval_nb_samtale_runtime": 131.6921,
|
166 |
+
"eval_nb_samtale_samples_per_second": 0.972,
|
167 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
168 |
+
"eval_nb_samtale_wer": 12.238493723849372,
|
169 |
+
"step": 126
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 1.43,
|
173 |
+
"eval_bigbrother_loss": 0.7359280586242676,
|
174 |
+
"eval_bigbrother_runtime": 252.6466,
|
175 |
+
"eval_bigbrother_samples_per_second": 1.013,
|
176 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
177 |
+
"eval_bigbrother_wer": 28.355555555555554,
|
178 |
+
"step": 126
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"epoch": 1.67,
|
182 |
+
"eval_rundkast_loss": 0.3938118815422058,
|
183 |
+
"eval_rundkast_runtime": 258.8717,
|
184 |
+
"eval_rundkast_samples_per_second": 0.989,
|
185 |
+
"eval_rundkast_steps_per_second": 0.015,
|
186 |
+
"eval_rundkast_wer": 13.967437874892887,
|
187 |
+
"step": 147
|
188 |
+
},
|
189 |
+
{
|
190 |
+
"epoch": 1.67,
|
191 |
+
"eval_nb_samtale_loss": 0.3065837025642395,
|
192 |
+
"eval_nb_samtale_runtime": 131.151,
|
193 |
+
"eval_nb_samtale_samples_per_second": 0.976,
|
194 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
195 |
+
"eval_nb_samtale_wer": 11.436541143654114,
|
196 |
+
"step": 147
|
197 |
+
},
|
198 |
+
{
|
199 |
+
"epoch": 1.67,
|
200 |
+
"eval_bigbrother_loss": 0.6836585998535156,
|
201 |
+
"eval_bigbrother_runtime": 253.257,
|
202 |
+
"eval_bigbrother_samples_per_second": 1.011,
|
203 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
204 |
+
"eval_bigbrother_wer": 27.644444444444442,
|
205 |
+
"step": 147
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 1.91,
|
209 |
+
"eval_rundkast_loss": 0.3802303373813629,
|
210 |
+
"eval_rundkast_runtime": 253.7316,
|
211 |
+
"eval_rundkast_samples_per_second": 1.009,
|
212 |
+
"eval_rundkast_steps_per_second": 0.016,
|
213 |
+
"eval_rundkast_wer": 13.753213367609254,
|
214 |
+
"step": 168
|
215 |
+
},
|
216 |
+
{
|
217 |
+
"epoch": 1.91,
|
218 |
+
"eval_nb_samtale_loss": 0.2869262993335724,
|
219 |
+
"eval_nb_samtale_runtime": 130.461,
|
220 |
+
"eval_nb_samtale_samples_per_second": 0.981,
|
221 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
222 |
+
"eval_nb_samtale_wer": 10.460251046025103,
|
223 |
+
"step": 168
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 1.91,
|
227 |
+
"eval_bigbrother_loss": 0.65956711769104,
|
228 |
+
"eval_bigbrother_runtime": 251.8276,
|
229 |
+
"eval_bigbrother_samples_per_second": 1.017,
|
230 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
231 |
+
"eval_bigbrother_wer": 27.68888888888889,
|
232 |
+
"step": 168
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 2.15,
|
236 |
+
"eval_rundkast_loss": 0.37286698818206787,
|
237 |
+
"eval_rundkast_runtime": 252.2645,
|
238 |
+
"eval_rundkast_samples_per_second": 1.015,
|
239 |
+
"eval_rundkast_steps_per_second": 0.016,
|
240 |
+
"eval_rundkast_wer": 13.581833761782347,
|
241 |
+
"step": 189
|
242 |
+
},
|
243 |
+
{
|
244 |
+
"epoch": 2.15,
|
245 |
+
"eval_nb_samtale_loss": 0.27960070967674255,
|
246 |
+
"eval_nb_samtale_runtime": 131.5439,
|
247 |
+
"eval_nb_samtale_samples_per_second": 0.973,
|
248 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
249 |
+
"eval_nb_samtale_wer": 10.355648535564853,
|
250 |
+
"step": 189
|
251 |
+
},
|
252 |
+
{
|
253 |
+
"epoch": 2.15,
|
254 |
+
"eval_bigbrother_loss": 0.650139331817627,
|
255 |
+
"eval_bigbrother_runtime": 252.878,
|
256 |
+
"eval_bigbrother_samples_per_second": 1.012,
|
257 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
258 |
+
"eval_bigbrother_wer": 27.644444444444442,
|
259 |
+
"step": 189
|
260 |
+
},
|
261 |
+
{
|
262 |
+
"epoch": 2.27,
|
263 |
+
"grad_norm": 1.7029398679733276,
|
264 |
+
"learning_rate": 2.0000000000000003e-06,
|
265 |
+
"loss": 0.3125,
|
266 |
+
"step": 200
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 2.39,
|
270 |
+
"eval_rundkast_loss": 0.37109890580177307,
|
271 |
+
"eval_rundkast_runtime": 260.7651,
|
272 |
+
"eval_rundkast_samples_per_second": 0.982,
|
273 |
+
"eval_rundkast_steps_per_second": 0.015,
|
274 |
+
"eval_rundkast_wer": 13.79605826906598,
|
275 |
+
"step": 210
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 2.39,
|
279 |
+
"eval_nb_samtale_loss": 0.2866850793361664,
|
280 |
+
"eval_nb_samtale_runtime": 134.0747,
|
281 |
+
"eval_nb_samtale_samples_per_second": 0.955,
|
282 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
283 |
+
"eval_nb_samtale_wer": 10.564853556485355,
|
284 |
+
"step": 210
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 2.39,
|
288 |
+
"eval_bigbrother_loss": 0.72154700756073,
|
289 |
+
"eval_bigbrother_runtime": 259.4669,
|
290 |
+
"eval_bigbrother_samples_per_second": 0.987,
|
291 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
292 |
+
"eval_bigbrother_wer": 26.93333333333333,
|
293 |
+
"step": 210
|
294 |
+
},
|
295 |
+
{
|
296 |
+
"epoch": 2.62,
|
297 |
+
"eval_rundkast_loss": 0.3417912721633911,
|
298 |
+
"eval_rundkast_runtime": 287.8177,
|
299 |
+
"eval_rundkast_samples_per_second": 0.889,
|
300 |
+
"eval_rundkast_steps_per_second": 0.014,
|
301 |
+
"eval_rundkast_wer": 15.424164524421593,
|
302 |
+
"step": 231
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 2.62,
|
306 |
+
"eval_nb_samtale_loss": 0.25085484981536865,
|
307 |
+
"eval_nb_samtale_runtime": 133.6313,
|
308 |
+
"eval_nb_samtale_samples_per_second": 0.958,
|
309 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
310 |
+
"eval_nb_samtale_wer": 9.867503486750348,
|
311 |
+
"step": 231
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 2.62,
|
315 |
+
"eval_bigbrother_loss": 0.5950784683227539,
|
316 |
+
"eval_bigbrother_runtime": 320.7203,
|
317 |
+
"eval_bigbrother_samples_per_second": 0.798,
|
318 |
+
"eval_bigbrother_steps_per_second": 0.012,
|
319 |
+
"eval_bigbrother_wer": 27.37777777777778,
|
320 |
+
"step": 231
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 2.86,
|
324 |
+
"eval_rundkast_loss": 0.3154206871986389,
|
325 |
+
"eval_rundkast_runtime": 258.2448,
|
326 |
+
"eval_rundkast_samples_per_second": 0.991,
|
327 |
+
"eval_rundkast_steps_per_second": 0.015,
|
328 |
+
"eval_rundkast_wer": 13.453299057412169,
|
329 |
+
"step": 252
|
330 |
+
},
|
331 |
+
{
|
332 |
+
"epoch": 2.86,
|
333 |
+
"eval_nb_samtale_loss": 0.23974575102329254,
|
334 |
+
"eval_nb_samtale_runtime": 135.359,
|
335 |
+
"eval_nb_samtale_samples_per_second": 0.946,
|
336 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
337 |
+
"eval_nb_samtale_wer": 9.518828451882845,
|
338 |
+
"step": 252
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 2.86,
|
342 |
+
"eval_bigbrother_loss": 0.5906115770339966,
|
343 |
+
"eval_bigbrother_runtime": 261.1086,
|
344 |
+
"eval_bigbrother_samples_per_second": 0.98,
|
345 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
346 |
+
"eval_bigbrother_wer": 26.8,
|
347 |
+
"step": 252
|
348 |
+
},
|
349 |
+
{
|
350 |
+
"epoch": 3.1,
|
351 |
+
"eval_rundkast_loss": 0.30801013112068176,
|
352 |
+
"eval_rundkast_runtime": 254.8501,
|
353 |
+
"eval_rundkast_samples_per_second": 1.005,
|
354 |
+
"eval_rundkast_steps_per_second": 0.016,
|
355 |
+
"eval_rundkast_wer": 13.410454155955442,
|
356 |
+
"step": 273
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 3.1,
|
360 |
+
"eval_nb_samtale_loss": 0.23249460756778717,
|
361 |
+
"eval_nb_samtale_runtime": 130.939,
|
362 |
+
"eval_nb_samtale_samples_per_second": 0.978,
|
363 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
364 |
+
"eval_nb_samtale_wer": 9.797768479776849,
|
365 |
+
"step": 273
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 3.1,
|
369 |
+
"eval_bigbrother_loss": 0.5573908090591431,
|
370 |
+
"eval_bigbrother_runtime": 253.2503,
|
371 |
+
"eval_bigbrother_samples_per_second": 1.011,
|
372 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
373 |
+
"eval_bigbrother_wer": 27.6,
|
374 |
+
"step": 273
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 3.34,
|
378 |
+
"eval_rundkast_loss": 0.3013758659362793,
|
379 |
+
"eval_rundkast_runtime": 252.5508,
|
380 |
+
"eval_rundkast_samples_per_second": 1.014,
|
381 |
+
"eval_rundkast_steps_per_second": 0.016,
|
382 |
+
"eval_rundkast_wer": 13.110539845758353,
|
383 |
+
"step": 294
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 3.34,
|
387 |
+
"eval_nb_samtale_loss": 0.22886353731155396,
|
388 |
+
"eval_nb_samtale_runtime": 130.5994,
|
389 |
+
"eval_nb_samtale_samples_per_second": 0.98,
|
390 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
391 |
+
"eval_nb_samtale_wer": 9.483960948396094,
|
392 |
+
"step": 294
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 3.34,
|
396 |
+
"eval_bigbrother_loss": 0.5597022771835327,
|
397 |
+
"eval_bigbrother_runtime": 252.6597,
|
398 |
+
"eval_bigbrother_samples_per_second": 1.013,
|
399 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
400 |
+
"eval_bigbrother_wer": 28.044444444444444,
|
401 |
+
"step": 294
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 3.41,
|
405 |
+
"grad_norm": 2.1695916652679443,
|
406 |
+
"learning_rate": 3e-06,
|
407 |
+
"loss": 0.2225,
|
408 |
+
"step": 300
|
409 |
+
},
|
410 |
+
{
|
411 |
+
"epoch": 3.58,
|
412 |
+
"eval_rundkast_loss": 0.29593202471733093,
|
413 |
+
"eval_rundkast_runtime": 265.0209,
|
414 |
+
"eval_rundkast_samples_per_second": 0.966,
|
415 |
+
"eval_rundkast_steps_per_second": 0.015,
|
416 |
+
"eval_rundkast_wer": 13.53898886032562,
|
417 |
+
"step": 315
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 3.58,
|
421 |
+
"eval_nb_samtale_loss": 0.23120403289794922,
|
422 |
+
"eval_nb_samtale_runtime": 137.2483,
|
423 |
+
"eval_nb_samtale_samples_per_second": 0.933,
|
424 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
425 |
+
"eval_nb_samtale_wer": 9.9721059972106,
|
426 |
+
"step": 315
|
427 |
+
},
|
428 |
+
{
|
429 |
+
"epoch": 3.58,
|
430 |
+
"eval_bigbrother_loss": 0.5614113807678223,
|
431 |
+
"eval_bigbrother_runtime": 257.5911,
|
432 |
+
"eval_bigbrother_samples_per_second": 0.994,
|
433 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
434 |
+
"eval_bigbrother_wer": 27.73333333333333,
|
435 |
+
"step": 315
|
436 |
+
},
|
437 |
+
{
|
438 |
+
"epoch": 3.82,
|
439 |
+
"eval_rundkast_loss": 0.2934091091156006,
|
440 |
+
"eval_rundkast_runtime": 253.624,
|
441 |
+
"eval_rundkast_samples_per_second": 1.009,
|
442 |
+
"eval_rundkast_steps_per_second": 0.016,
|
443 |
+
"eval_rundkast_wer": 13.453299057412169,
|
444 |
+
"step": 336
|
445 |
+
},
|
446 |
+
{
|
447 |
+
"epoch": 3.82,
|
448 |
+
"eval_nb_samtale_loss": 0.2249545007944107,
|
449 |
+
"eval_nb_samtale_runtime": 134.2615,
|
450 |
+
"eval_nb_samtale_samples_per_second": 0.953,
|
451 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
452 |
+
"eval_nb_samtale_wer": 9.693165969316597,
|
453 |
+
"step": 336
|
454 |
+
},
|
455 |
+
{
|
456 |
+
"epoch": 3.82,
|
457 |
+
"eval_bigbrother_loss": 0.5466703176498413,
|
458 |
+
"eval_bigbrother_runtime": 258.8204,
|
459 |
+
"eval_bigbrother_samples_per_second": 0.989,
|
460 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
461 |
+
"eval_bigbrother_wer": 27.68888888888889,
|
462 |
+
"step": 336
|
463 |
+
},
|
464 |
+
{
|
465 |
+
"epoch": 4.06,
|
466 |
+
"eval_rundkast_loss": 0.2969070076942444,
|
467 |
+
"eval_rundkast_runtime": 253.4479,
|
468 |
+
"eval_rundkast_samples_per_second": 1.01,
|
469 |
+
"eval_rundkast_steps_per_second": 0.016,
|
470 |
+
"eval_rundkast_wer": 13.753213367609254,
|
471 |
+
"step": 357
|
472 |
+
},
|
473 |
+
{
|
474 |
+
"epoch": 4.06,
|
475 |
+
"eval_nb_samtale_loss": 0.2252650409936905,
|
476 |
+
"eval_nb_samtale_runtime": 133.6675,
|
477 |
+
"eval_nb_samtale_samples_per_second": 0.958,
|
478 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
479 |
+
"eval_nb_samtale_wer": 9.902370990237099,
|
480 |
+
"step": 357
|
481 |
+
},
|
482 |
+
{
|
483 |
+
"epoch": 4.06,
|
484 |
+
"eval_bigbrother_loss": 0.5490403771400452,
|
485 |
+
"eval_bigbrother_runtime": 252.1606,
|
486 |
+
"eval_bigbrother_samples_per_second": 1.015,
|
487 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
488 |
+
"eval_bigbrother_wer": 27.51111111111111,
|
489 |
+
"step": 357
|
490 |
+
},
|
491 |
+
{
|
492 |
+
"epoch": 4.3,
|
493 |
+
"eval_rundkast_loss": 0.29715079069137573,
|
494 |
+
"eval_rundkast_runtime": 263.3468,
|
495 |
+
"eval_rundkast_samples_per_second": 0.972,
|
496 |
+
"eval_rundkast_steps_per_second": 0.015,
|
497 |
+
"eval_rundkast_wer": 13.367609254498714,
|
498 |
+
"step": 378
|
499 |
+
},
|
500 |
+
{
|
501 |
+
"epoch": 4.3,
|
502 |
+
"eval_nb_samtale_loss": 0.22843754291534424,
|
503 |
+
"eval_nb_samtale_runtime": 136.3747,
|
504 |
+
"eval_nb_samtale_samples_per_second": 0.939,
|
505 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
506 |
+
"eval_nb_samtale_wer": 9.797768479776849,
|
507 |
+
"step": 378
|
508 |
+
},
|
509 |
+
{
|
510 |
+
"epoch": 4.3,
|
511 |
+
"eval_bigbrother_loss": 0.5757928490638733,
|
512 |
+
"eval_bigbrother_runtime": 252.9905,
|
513 |
+
"eval_bigbrother_samples_per_second": 1.012,
|
514 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
515 |
+
"eval_bigbrother_wer": 27.37777777777778,
|
516 |
+
"step": 378
|
517 |
+
},
|
518 |
+
{
|
519 |
+
"epoch": 4.53,
|
520 |
+
"eval_rundkast_loss": 0.2953430414199829,
|
521 |
+
"eval_rundkast_runtime": 258.4676,
|
522 |
+
"eval_rundkast_samples_per_second": 0.99,
|
523 |
+
"eval_rundkast_steps_per_second": 0.015,
|
524 |
+
"eval_rundkast_wer": 13.453299057412169,
|
525 |
+
"step": 399
|
526 |
+
},
|
527 |
+
{
|
528 |
+
"epoch": 4.53,
|
529 |
+
"eval_nb_samtale_loss": 0.22929418087005615,
|
530 |
+
"eval_nb_samtale_runtime": 135.2614,
|
531 |
+
"eval_nb_samtale_samples_per_second": 0.946,
|
532 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
533 |
+
"eval_nb_samtale_wer": 9.588563458856346,
|
534 |
+
"step": 399
|
535 |
+
},
|
536 |
+
{
|
537 |
+
"epoch": 4.53,
|
538 |
+
"eval_bigbrother_loss": 0.5682810544967651,
|
539 |
+
"eval_bigbrother_runtime": 261.904,
|
540 |
+
"eval_bigbrother_samples_per_second": 0.977,
|
541 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
542 |
+
"eval_bigbrother_wer": 27.68888888888889,
|
543 |
+
"step": 399
|
544 |
+
},
|
545 |
+
{
|
546 |
+
"epoch": 4.55,
|
547 |
+
"grad_norm": 1.9239732027053833,
|
548 |
+
"learning_rate": 4.000000000000001e-06,
|
549 |
+
"loss": 0.1597,
|
550 |
+
"step": 400
|
551 |
+
},
|
552 |
+
{
|
553 |
+
"epoch": 4.77,
|
554 |
+
"eval_rundkast_loss": 0.2963680922985077,
|
555 |
+
"eval_rundkast_runtime": 257.7303,
|
556 |
+
"eval_rundkast_samples_per_second": 0.993,
|
557 |
+
"eval_rundkast_steps_per_second": 0.016,
|
558 |
+
"eval_rundkast_wer": 13.496143958868895,
|
559 |
+
"step": 420
|
560 |
+
},
|
561 |
+
{
|
562 |
+
"epoch": 4.77,
|
563 |
+
"eval_nb_samtale_loss": 0.22047586739063263,
|
564 |
+
"eval_nb_samtale_runtime": 134.462,
|
565 |
+
"eval_nb_samtale_samples_per_second": 0.952,
|
566 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
567 |
+
"eval_nb_samtale_wer": 9.309623430962342,
|
568 |
+
"step": 420
|
569 |
+
},
|
570 |
+
{
|
571 |
+
"epoch": 4.77,
|
572 |
+
"eval_bigbrother_loss": 0.5600541830062866,
|
573 |
+
"eval_bigbrother_runtime": 259.0585,
|
574 |
+
"eval_bigbrother_samples_per_second": 0.988,
|
575 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
576 |
+
"eval_bigbrother_wer": 27.200000000000003,
|
577 |
+
"step": 420
|
578 |
+
},
|
579 |
+
{
|
580 |
+
"epoch": 5.01,
|
581 |
+
"eval_rundkast_loss": 0.2924000322818756,
|
582 |
+
"eval_rundkast_runtime": 253.711,
|
583 |
+
"eval_rundkast_samples_per_second": 1.009,
|
584 |
+
"eval_rundkast_steps_per_second": 0.016,
|
585 |
+
"eval_rundkast_wer": 13.281919451585262,
|
586 |
+
"step": 441
|
587 |
+
},
|
588 |
+
{
|
589 |
+
"epoch": 5.01,
|
590 |
+
"eval_nb_samtale_loss": 0.2149951457977295,
|
591 |
+
"eval_nb_samtale_runtime": 134.7407,
|
592 |
+
"eval_nb_samtale_samples_per_second": 0.95,
|
593 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
594 |
+
"eval_nb_samtale_wer": 9.483960948396094,
|
595 |
+
"step": 441
|
596 |
+
},
|
597 |
+
{
|
598 |
+
"epoch": 5.01,
|
599 |
+
"eval_bigbrother_loss": 0.5668565630912781,
|
600 |
+
"eval_bigbrother_runtime": 259.6443,
|
601 |
+
"eval_bigbrother_samples_per_second": 0.986,
|
602 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
603 |
+
"eval_bigbrother_wer": 26.8,
|
604 |
+
"step": 441
|
605 |
+
},
|
606 |
+
{
|
607 |
+
"epoch": 5.25,
|
608 |
+
"eval_rundkast_loss": 0.31433504819869995,
|
609 |
+
"eval_rundkast_runtime": 258.8641,
|
610 |
+
"eval_rundkast_samples_per_second": 0.989,
|
611 |
+
"eval_rundkast_steps_per_second": 0.015,
|
612 |
+
"eval_rundkast_wer": 13.79605826906598,
|
613 |
+
"step": 462
|
614 |
+
},
|
615 |
+
{
|
616 |
+
"epoch": 5.25,
|
617 |
+
"eval_nb_samtale_loss": 0.2276706099510193,
|
618 |
+
"eval_nb_samtale_runtime": 135.9916,
|
619 |
+
"eval_nb_samtale_samples_per_second": 0.941,
|
620 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
621 |
+
"eval_nb_samtale_wer": 9.344490934449093,
|
622 |
+
"step": 462
|
623 |
+
},
|
624 |
+
{
|
625 |
+
"epoch": 5.25,
|
626 |
+
"eval_bigbrother_loss": 0.6108298301696777,
|
627 |
+
"eval_bigbrother_runtime": 258.4975,
|
628 |
+
"eval_bigbrother_samples_per_second": 0.99,
|
629 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
630 |
+
"eval_bigbrother_wer": 28.48888888888889,
|
631 |
+
"step": 462
|
632 |
+
},
|
633 |
+
{
|
634 |
+
"epoch": 5.49,
|
635 |
+
"eval_rundkast_loss": 0.31115537881851196,
|
636 |
+
"eval_rundkast_runtime": 254.7658,
|
637 |
+
"eval_rundkast_samples_per_second": 1.005,
|
638 |
+
"eval_rundkast_steps_per_second": 0.016,
|
639 |
+
"eval_rundkast_wer": 13.710368466152529,
|
640 |
+
"step": 483
|
641 |
+
},
|
642 |
+
{
|
643 |
+
"epoch": 5.49,
|
644 |
+
"eval_nb_samtale_loss": 0.2282358556985855,
|
645 |
+
"eval_nb_samtale_runtime": 134.7289,
|
646 |
+
"eval_nb_samtale_samples_per_second": 0.95,
|
647 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
648 |
+
"eval_nb_samtale_wer": 9.06555090655509,
|
649 |
+
"step": 483
|
650 |
+
},
|
651 |
+
{
|
652 |
+
"epoch": 5.49,
|
653 |
+
"eval_bigbrother_loss": 0.6029994487762451,
|
654 |
+
"eval_bigbrother_runtime": 253.4743,
|
655 |
+
"eval_bigbrother_samples_per_second": 1.01,
|
656 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
657 |
+
"eval_bigbrother_wer": 27.066666666666666,
|
658 |
+
"step": 483
|
659 |
+
},
|
660 |
+
{
|
661 |
+
"epoch": 5.68,
|
662 |
+
"grad_norm": 2.697171449661255,
|
663 |
+
"learning_rate": 5e-06,
|
664 |
+
"loss": 0.1207,
|
665 |
+
"step": 500
|
666 |
+
},
|
667 |
+
{
|
668 |
+
"epoch": 5.73,
|
669 |
+
"eval_rundkast_loss": 0.31634020805358887,
|
670 |
+
"eval_rundkast_runtime": 260.8574,
|
671 |
+
"eval_rundkast_samples_per_second": 0.981,
|
672 |
+
"eval_rundkast_steps_per_second": 0.015,
|
673 |
+
"eval_rundkast_wer": 13.838903170522707,
|
674 |
+
"step": 504
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 5.73,
|
678 |
+
"eval_nb_samtale_loss": 0.22956953942775726,
|
679 |
+
"eval_nb_samtale_runtime": 134.4632,
|
680 |
+
"eval_nb_samtale_samples_per_second": 0.952,
|
681 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
682 |
+
"eval_nb_samtale_wer": 9.379358437935844,
|
683 |
+
"step": 504
|
684 |
+
},
|
685 |
+
{
|
686 |
+
"epoch": 5.73,
|
687 |
+
"eval_bigbrother_loss": 0.6007498502731323,
|
688 |
+
"eval_bigbrother_runtime": 259.5455,
|
689 |
+
"eval_bigbrother_samples_per_second": 0.986,
|
690 |
+
"eval_bigbrother_steps_per_second": 0.015,
|
691 |
+
"eval_bigbrother_wer": 28.000000000000004,
|
692 |
+
"step": 504
|
693 |
+
},
|
694 |
+
{
|
695 |
+
"epoch": 5.97,
|
696 |
+
"eval_rundkast_loss": 0.3192276954650879,
|
697 |
+
"eval_rundkast_runtime": 253.8269,
|
698 |
+
"eval_rundkast_samples_per_second": 1.009,
|
699 |
+
"eval_rundkast_steps_per_second": 0.016,
|
700 |
+
"eval_rundkast_wer": 13.53898886032562,
|
701 |
+
"step": 525
|
702 |
+
},
|
703 |
+
{
|
704 |
+
"epoch": 5.97,
|
705 |
+
"eval_nb_samtale_loss": 0.22917582094669342,
|
706 |
+
"eval_nb_samtale_runtime": 134.9907,
|
707 |
+
"eval_nb_samtale_samples_per_second": 0.948,
|
708 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
709 |
+
"eval_nb_samtale_wer": 9.06555090655509,
|
710 |
+
"step": 525
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 5.97,
|
714 |
+
"eval_bigbrother_loss": 0.6029475331306458,
|
715 |
+
"eval_bigbrother_runtime": 253.265,
|
716 |
+
"eval_bigbrother_samples_per_second": 1.011,
|
717 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
718 |
+
"eval_bigbrother_wer": 26.666666666666668,
|
719 |
+
"step": 525
|
720 |
+
},
|
721 |
+
{
|
722 |
+
"epoch": 6.2,
|
723 |
+
"eval_rundkast_loss": 0.3402000665664673,
|
724 |
+
"eval_rundkast_runtime": 255.0851,
|
725 |
+
"eval_rundkast_samples_per_second": 1.004,
|
726 |
+
"eval_rundkast_steps_per_second": 0.016,
|
727 |
+
"eval_rundkast_wer": 13.367609254498714,
|
728 |
+
"step": 546
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 6.2,
|
732 |
+
"eval_nb_samtale_loss": 0.23988915979862213,
|
733 |
+
"eval_nb_samtale_runtime": 133.3224,
|
734 |
+
"eval_nb_samtale_samples_per_second": 0.96,
|
735 |
+
"eval_nb_samtale_steps_per_second": 0.015,
|
736 |
+
"eval_nb_samtale_wer": 8.647140864714087,
|
737 |
+
"step": 546
|
738 |
+
},
|
739 |
+
{
|
740 |
+
"epoch": 6.2,
|
741 |
+
"eval_bigbrother_loss": 0.6717787384986877,
|
742 |
+
"eval_bigbrother_runtime": 257.5221,
|
743 |
+
"eval_bigbrother_samples_per_second": 0.994,
|
744 |
+
"eval_bigbrother_steps_per_second": 0.016,
|
745 |
+
"eval_bigbrother_wer": 26.711111111111112,
|
746 |
+
"step": 546
|
747 |
+
}
|
748 |
+
],
|
749 |
+
"logging_steps": 100,
|
750 |
+
"max_steps": 1320,
|
751 |
+
"num_input_tokens_seen": 0,
|
752 |
+
"num_train_epochs": 15,
|
753 |
+
"save_steps": 21,
|
754 |
+
"total_flos": 8.900086563864576e+19,
|
755 |
+
"train_batch_size": 48,
|
756 |
+
"trial_name": null,
|
757 |
+
"trial_params": null
|
758 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0af81cda6f4e6e8ca7298eea8143a7256eb99e6fe1fe28d8f057758dad120caf
|
3 |
+
size 5112
|